During our shipboard paleomagnetic studies for Leg 183, we characterized the stable components of NRM through thermal and AF demagnetization. Details are illustrated and described in the site chapters of the Leg 183 Initial Reports volume (Coffin, Frey, Wallace, et al. 2000). Here, we only highlight the most relevant points in terms of demagnetization behavior, paleolatitude, and magnetic polarity.
A stable component of magnetization in the basement rocks from Sites 1136, 1137, 1138, 1139, and 1140 was revealed. In most samples, this stable component of magnetization was straightforward upon AF or thermal treatments. A generally minor secondary component, probably a viscous overprint from the recent field, was present in some samples but was easily removed at peak fields of 20 mT or at a low temperature of ~150ºC. The median destructive fields ranged mostly in the 30- to 50-mT interval, suggesting the existence of small pseudosingle grains as remanence carriers (Dunlop and Özdemir, 1997). Most samples have unblocking temperatures of ~580ºC, suggesting titanium-poor titanomagnetite is responsible for the remanent magnetization. In some samples, we observed two magnetic phases during stepwise thermal demagnetization and susceptibility measurements. A high-temperature phase is characterized by an unblocking temperature of ~580ºC, and a low-temperature phase is characterized by an unblocking temperature of ~300ºC. In a few samples, the stable component was not unblocked until 620ºC. Magnetic susceptibilities for these samples increased at 400°-500ºC and decreased at 500°-620ºC, indicating changes from (titano)maghemite to (titano)magnetite and then to (titano)maghematite, respectively.
The magnetically cleaned mean inclinations at Sites 1136 and 1137 correspond to paleolatitudes that are ~4º-10º farther north than present-day site latitudes. Assuming the inclination represents the primary remanence at the time when these rocks were formed, the paleolatitude results would suggest that the southern Kerguelen Plateau and Elan Bank have moved southward since Cretaceous time. This southward movement seems to be consistent with previous paleomagnetic results from the southern Kerguelen Plateau, which show a difference of 13º-16º between present day latitude and paleolatitude of basement rocks (Inokuchi and Heider, 1992).
The ChRM for Sites 1136 (119 Ma), 1137 (108 Ma), 1138 (100 Ma), 1141 (95 Ma), and 1142 (95 Ma) are all normal, which may indicate the magnetization was acquired during the Cretaceous Normal Superchron (121-83 Ma) and is also consistent with the ages of the rocks. However, there is no independent proof of this timing, as no meaningful fold tests are available on the basement rocks. Thus, the normal magnetization could be an overprint resulting in the acquisition of secondary magnetization, which is now accepted as a widespread phenomenon (e.g., Van der Voo, 1993). Interestingly, preliminary shipboard paleomagnetic studies on six samples from Site 1139 (68 Ma) showed reversed polarity magnetization. There is also a magnetic reversal zone at the boundary between basement rock Units 1 and 2 at Site 1140 (34 Ma) (Shipboard Scientific Party, 2000b). The main issues that are important to consider are the timing of the reversal zones in the basalt (i.e., whether it is the same as the formation of the basalts) and the ability of basalts to carry the remanent magnetization (i.e., whether it is stable over geologically significant time intervals). Additional rock magnetic data are needed to constrain the observed reversal polarity zone.
Curie temperature determinations of samples from Sites 1136-1142 are presented in Table T2. Strong field thermomagnetic curves were obtained to determine the magnetic phases in the samples. According to Curie temperatures (Table T2), three different groups can be recognized from the Leg 183 samples. Group l (Fig. 2A) is characterized by a single ferromagnetic phase with Curie temperatures between 480º and 580ºC, compatible with that of Ti-poor titanomagnetites. The cooling and heating curves are reasonably reversible. Most subaerial flows from Sites 1136, 1137, 1138, 1139, 1141, and 1142 belong to this group (Table T2). Group 2 has lower Curie temperatures (165º-359ºC) (see Table T2) that are typical of titanium-rich titanomagnetite (such as TM60) or low-temperature oxidized titanomaghemites. Group 2 curves (Fig. 2B) were mainly observed for pillow basalts from Site 1140. Other rock samples that can be included in this group are from the lower parts of Sites 1136 and 1138, which exhibit similar low Curie temperatures. Samples 183-1136A-19R-1, 122-124 cm; 183-1138A-88R-1, 53-55 cm; and 183-1138A-89R-3, 40-42 cm, are exceptions (Table T2). Samples belonging to Group 3 have multiple magnetic phases and come mainly from Site 1139. For the purpose of showing the entire suite of measurements from one representative sample, we choose to show Sample 183-1138A-85R-2, 89-91 cm, which displays the same behavior. The irreversible thermomagnetic curve of this sample displays one magnetic phase with Curie temperature of ~380ºC on heating (Fig. 2C). The second highest Curie temperature phase is observed at ~540ºC. The large difference between heating and cooling of the sample suggests that a low-temperature oxidized titanomagnetite is the main magnetic mineral.
Comparison of the two Curie temperatures (obtained by low-field continuous susceptibility and high-field magnetic moment vs. temperature runs) reveals that, in general, Curie temperature by susceptibility as a function of temperature is less than those determined by high-field magnetic moment runs (Table T2). An extreme discrepancy in Curie temperatures is presented by Sample 183-1140A-37R-4, 16-18 (Table T2). The strong-field Curie temperature is found to be 540ºC, which is consistent with titanomagnetite 05, but the susceptibility data are flat after 230ºC. It is not clear whether this is due to nonhomogeneous sub-samples used in the two instruments. Thermomagnetic curves for the two crystal vitric tuff samples (Table T2) from Site 1137 show behavior that cannot be interpreted in a simple manner. The magnetic mineralogy for these two samples is complex, and the results are not readily explained at present.
The samples analyzed in this study indicate that subaerial basalt samples from Sites 1136, 1137, 1138, and 1141 show the predominance of the PSD size, probably indicating a mixture of MD and SD grains (Fig. F3). We note that if some superparamagnetic grains are also present, the measured coercive force and saturation magnetization may be somewhat lower and larger, respectively. The magnetic grain sizes of the pillow basalt samples from Site 1140 fall near the boundary between SD and PSD. In contrast, samples from Sites 1139 and 1142 distribute toward the boundary between PSD and MD, with a few samples even falling in the MD size (Fig. F3; Table T3). All the samples from Site 1139 also showed constricted hysteresis loops (or "wasp-waisted" loops), indicating multiple magnetic phases. The corresponding hysteresis ratios (Fig. F3; Table T3) appear to suggest the secondary magnetite formation reported by various colleagues (e.g., Channel and McCabe, 1994). Examples of a room-temperature hysteresis loop from the Kerguelen Plateau for representative basalt samples that exhibit SD, MD, and PSD behavior are shown in Figure F4.
For a selected group of samples, we also examined the change of hysteresis loops as a function of temperature (10-400 K) to detect changes in the domain state at low or high temperature. Figure F5 shows selected hysteresis loops for two representative samples. The nonsaturating linear part in the subaerial basalt sample (Sample 183-1136A-15R-3, 109-111 cm) indicates a paramagnetic component (Fig. 5A). After that, loop parameters for this sample show little temperature variant. On the other hand, the fine-grained pillow basalt sample (Sample 183-1140A-27R-5, 41-43 cm) displayed a different domain structure (Fig. F5B). In this sample, coercivity and saturation remanence measured at each temperature from the hysteresis loop exhibit similar dependences on temperature for a TM60 sample from the Mogo Hill breccia (M. Jackson, pers. comm., 2000), with a rapid drop between 10 and 80 K. It is interesting to note that coercivity increases from 10 to 50 K then decreases at higher temperatures (Fig. F5B). Hysteresis loops for Sample 183-1139A-71R-4, 8-9 cm, maintained superparamagnetic loops throughout the entire temperature interval.
As shown in Figure F6, the low-temperature curves of SIRM both in zero field warming and cooling display a variety of features. These include an unblocking temperature in the vicinity of 40-50 K, which is probably caused by superparamagnetic magnetite particles (Moskowitz et al., 1993) and a decrease in remanence in the 100-120 K range, which is most likely caused by the Verwey transition (Verwey et al., 1947). Figure F6A shows cooling and warming curves for Sample 183-1136A-16R-1, 134-136 cm, which has PSD grain size. Remanence is lost at ~110-120 K as the sample cools and warms through the Verwey transition. On the other hand, no obvious Verwey transition is observed for the pillow basalt sample (Sample 183-1140A-27R-5, 41-43 cm) during cooling to 19 K. Upon warming from 19 K, however, a distinctive bend in remanence is present near 100 K. With continued warming, the decay of remanence is almost linear all the way to room temperature (Fig. F6B).
Figure F6C shows a third behavior. Sample 183-1138A-85R-2, 81-91 cm, has smaller grain size as indicated by hysteresis measurement. A rapid decrease in remanence between 10 and 40 K is observed (Fig. F6C). A similar phenomenon was observed in the study of oxidized synthetic magnetite by Özdemir et al. (1993) and was attributed to the presence of an ultra fine-grained superparamagnetic phase with a very low unblocking temperature. No pronounced remanence transition is observed, although the sample displays a more rapid decrease of remanence between 50 and 120 K. In comparison with the classic Verwey transition, the remanence transition for this sample is perhaps blurred over a broad temperature interval and is shifted toward lower temperatures. The low-temperature magnetometry is one of the major lines of evidence for the division of these three different groups.
Figures F7 and F8 show examples of the temperature dependence of magnetic susceptibility between 15 and 300 K for two Leg 183 core samples. The frequency-dependent susceptibility curves for the subaerial basalt sample (Sample 183-1141A-24R-2, 6-8 cm) from Site 1141 (Broken Ridge) are generally similar to those of synthetic titanomagnetite 28 and 41 (TM28 and TM41) (Jackson et al., 1998), with some frequency dependence from 25 to 100 K. An abrupt decrease of in-phase susceptibility (m´) is observed at ~40 K. Susceptibility then increases through the Verwey transition to 120 K. There is an abrupt frequency dependence increase between 55 and 95 K (see the lower part of the diagram in Fig. F7), a feature that is not yet understood. The frequency dependence of quadrature (m'') susceptibility observed from 100 to 200 K may indicate the presence of superparamagnetic grains. Interestingly, the fine-grained pillow basalt sample (Sample 183-1140A-37R-4, 16-18 cm) from Site 1140 (northern Kerguelen Plateau) shows very different frequency-dependent susceptibility curves (Fig. F8). There is no detectable frequency dependence of susceptibility until after 250 K, probably because the susceptibility is a superposition of paramagnetic and ferrimagnetic susceptibilities from different Fe-bearing phases.
The Mössbauer spectra obtained for the two samples that reflect the differences in the Leg 183 lithologies are shown in Figure F9. The spectrum of a subaerial basalt sample (Sample 183-1136A-16R-2, 140-142 cm) from Site 1136 (the Kerguelen Plateau) is shown in Figure 9A. At room temperature, the spectrum is dominated by three magnetically ordered components that have been tentatively identified as magnetite, maghemite, and hematite, respectively. The magnetite spectrum contains two components, one corresponding to the A site (Fe3+) and the other to the B site. Figure 9B shows the spectrum of a subaerial basalt sample from the Broken Ridge area (Sample 183-1141A-24R-2, 6-8 cm) with fitted curves. The values of the fitting parameters for this spectrum suggest two components that can be identified as maghemite and magnetite, respectively.